Imagine it, build it. It's shocking how simple it is to build complex workflows with this.
Things this has already done for me:
- Daily sentiment analysis on my products
- Regular podcasts on the most important AI news
- Prototyping new ideas
All with very short prompts!
Introducing Perplexity Computer.
Computer unifies every current AI capability into one system.
It can research, design, code, deploy, and manage any project end-to-end.
Today we're announcing that hybrid agentic inference is coming to Perplexity Computer.
Computer can split tasks between a local model running on your machine and frontier models in the cloud. This keeps private data on your device and maximizes token efficiency.
Coming soon.
Introducing Search as Code, our new search architecture for AI agents.
It writes Python that calls our search stack directly, instead of looping through function calls one at a time.
Available in the Perplexity Agent API, and now default in Computer.
https://t.co/ut6GGWQTVO
Having used the new app for the last month or so, I can tell you, I don't need the old one any ways, have gotten so used to the new one.
What I really like is the connection between the Perplexity app on my phone through Computer to Perplexity Personal Computer - IMO, much cleaner than controlling through something like Telegram.
QoL improvements to Perplexity Computer's ability to act as your personal CFO:
- You can now connect up to 30 accounts (prev. 5)
- We will now fetch up to 500 assets per account (prev. 200)
- You can set your annual income and risk tolerance, which will flavor answers from Computer about budget planning, portfolio management, etc.
Today we’re rolling out a new native Comet experience for iPad.
Comet now works naturally with iPadOS features like multiple windows and Split View, so you can work with Comet alongside the apps you already use.
Available now on the App Store.
Seamless access to your local files and apps unlocks completely new super powers with Computer.
- Take your unruly Apple Notes and turn them into your next prototype with quick commands.
- Search through a large folder of historical files and pull just the right context for a draft patent application.
- Take a series of messages from your partner and turn them into a clean grocery list 😅
Give it a try!
Today we're releasing Personal Computer.
Personal Computer integrates with the Perplexity Mac App for secure orchestration across your local files, native apps, and browser.
We’re rolling this out to all Perplexity Max subscribers and everyone on the waitlist starting today.
Perplexity started as a small business tool for ourselves. We had 4 people and no revenue with AI at our fingertips.
The pivot to Computer is actually a full circle. Founders are using it to grow companies that matter to the economy and their communities.
It’s rewarding to see it now powering small businesses and startups in big ways.
Perplexity is still a startup. We just 5X’ed revenue from $100M to $500M with only 34% growth in team size. 2x revenue growth in 2026 with same small team. And we’re just warming up.
Everyone here works at a small business, and everything we build is for people who build.
1. The labs lack high switching costs right now. Switching models is far to easy. If GPT 6 is 5% better than Opus 5, we all switch in 30s.
2. Tools have higher switching costs. You learn how to use them, introduce them to your teammates, leverage their differentiated plugin systems. Then, if GPT 6 is 5% better, you don't switch because Opus 5 is good enough and you love Claude Code.
3. Now the lab is no longer a simple commodity provider and they can start to slow down investment in training and they don't have to give you the best model all the time. They can keep it, or give it only to massive companies, etc. (Vs. if you were using opencode, and you could switch to GPT in 30s, they would have to keep competing on true model strength vs. price so you don't switch to GPT).
4. You use CC instead even though it's a little worse b/c the tokens are so much cheaper--$200/mo vs. $1500 a month. Such a great deal! There's the loss leadering.
5. This transforms the game from a product race into a fundraising race. The labs will win this, they have way more capital than anyone else.
6. Once the labs consolidate all the users, all the other product companies are forced sell themselves b/c the negative cash flow means they can no longer raise capital (probably to labs/big clouds), you raise the prices anyway.
Now you have no competition to provide the best model for the best price to everyone. And you also have less pressure to improve your product.
This is a very old game plan in silicon valley. Give it away for cheap or free and turn it into a capital race. Consumers always go for it--it's hard to resist an unsustainable deal. Then later on say "how did we end up here??" later when the prices are higher and products are worse.
Had a great time working on this with @charujswl! There’s really an excellent team on Realtime @OpenAI.
Definitely reach out if parts of this were useful when building your voice experience! Would love to bounce other ideas and share what we’re thinking about @perplexity_ai.
@cleondesigns@scspeier@Linds_Kado Yea for sure. I've been eng on the other side of this in the past :-) Products are always built organically in code, what ships is the reality, now that designers can actually code they can bend what ships back to the desired outcome when the eng didn't quite get it right.
@cleondesigns@scspeier@Linds_Kado What do you mean by design doc here? In this world, the code or PR becomes the source of truth. If the PR evolves it's design meeting reality and becoming the embodiment of the idea.